1. Field of the Invention
The present invention relates to the chemical vapor deposition (CVD) formation of Pb(Zr,Ti)O.sub.3 materials modified with Group II cations (Sr, Ca, Ba and/or Mg) on the A-sites thereof, and Nb and/or Ta on the B-sites thereof, and to (Pb, Sr, Ca, Ba, Mg)(Zr, Ti, Nb, Ta)O.sub.3 films having utility in ferroelectric random access memories, high performance thin film microactuators, and in related device applications.
2. Description of the Related Art
Ferroelectric materials are presently finding increased application in devices including non-volatile ferroelectric random access memories (NV-FeRAMs), uncooled infrared (IR) detectors, spatial light modulators, and microelectromechanical systems (MEMS). Many of these applications require optimized ferroelectric, pyroelectric and related properties, which are known to be sensitive to film compositions and incorporations of dopants or modifiers.
In order to effect such compositional variation, there is a need in the art for corresponding processes enabling the production of perovskite films with superior compositional and performance properties.
Directing the discussion now to a relevant background aspect of the present invention, the development of reliable actuation methods and devices is one of the key challenges in the evolution from micromachined sensors to microelectromechanical systems (MEMS). High quality piezoelectric films possess numerous properties of technological importance for such MEMS applications, including high electromechanical coupling coefficients and high piezoelectric coefficients. The most common family of material exhibiting both of these characteristics are based on Pb(Zr.sub.y Ti.sub.1-y)O.sub.3 (PZT).
One primary factor limiting development of piezoelectric MEMS has been the lack of suitable, high quality thin film piezoelectric materials. PZT and related compositions are the best piezoelectric materials available in bulk form and are a logical choice for thin-film microactuator applications.
A number of microactuator devices can be envisioned that are based on cantilever-type deflection, including optical devices and liquid control devices. Depending upon the application, the requirements far operating deflection needed in such devices may vary widely. In a cantilevered piezoelectric microactuator of the type that may be usefully employed in positioners and microvalves, the achievable deflection for an applied voltage is directly proportional to the piezoelectric coefficient d.sub.31. Since the deflection is directly proportional to the applied voltage and to the piezoelectric coefficient d.sub.31, an increase in d.sub.31 at a given voltage increases the deflection. Looking at this relationship another way, for a given deflection, the drive voltage is reduced with increased d.sub.31. Lower drive voltage requirements are also a significant advantage as integration of PZT microactuators into integrated circuit (IC) devices is contemplated by the art and this remains important even for small displacement devices.
Accordingly, it would be a substantial advance in the art to develop compositions and process technology for the achievement of high quality films that are able to maximize deflection for a given drive voltage in microactuator applications. Similarly, these high quality films may have other advantageous properties.
Integration of thin film PZT and related materials into MEMS applications requires a well-controlled process that affords precise control of composition to maintain acceptable device performance across a wafer and from wafer-to-wafer. In addition, good step coverage is required for micromachining of the devices to protect the edges of features from unwanted etching. Finally, the process must be highly economical. This last requirement is comprised of several factors including the ability to process large area Si wafers and achieve high process throughput. Although the state of the art of bulk piezoelectric ceramic materials has changed little in the last decade, considerable effort has been focused on techniques to produce thin films of PZT and related materials.
RF sputtering ("Epitaxial Growth and Electrical Properties of Ferroelectric Pb(Zr.sub.0.9 Ti.sub.0.1)O.sub.3 Films by Reactive Sputtering," T. Okamura, M. Adachi, T. Shiosaki, A. Kawabata, Jap. J. Appl. Phys 30-1 (1991): 1034), sol-gel formation ("Low Temperature Perovskite Formation of Lead Zirconate Titanate Thin Films by a Seeding Process," C. K. Wok and S. B Desu, J. Mater. Res. 8 (1993): 339), and CVD ("Preparation and Properties of (Pb,La)(Zr,Ti)O.sub.3 Thin Films by Metalorganic Chemical Vapor Deposition," M. Okada and K. Tominaga, J. Appl. Phys. 71 (1992): 1955; and "Growth and Characterization of Ferroelectric Pb(Zr,Ti)O.sub.3 Thin Films by MOCVD Using a 6 Inch Single Wafer CVD System," M. Shimizu, M. Fujimoto, T. Katayama, T. Shiosaki, K. Nakaya, M. Fukagawa, and E. Tanikawa, ISIF'93 Proceedings, Colorado, Springs, Colo. (1993)) have all been used to make high quality thin film PZT.
RF sputtering is an inherently low deposition rate process for complex oxide materials like PZT and uniform composition is difficult to achieve across large areas. In addition, as sputtering targets wear, composition can drift and cross-target contamination is extremely problematic for process control. Sol-gel processes offer better control of composition, but have poor step coverage. Moreover, sol-gel processing of PZT requires post-deposition annealing, which can lead to vaporization and loss of Pb, and can affect underlying IC structures. Although progress has been made in lowering processing temperatures, for example by the use of seed layers ("Low Temperature Perovskite Formation of Lead Zirconate Titanate Thin Films by a Seeding Process," C. K. Kwok and S. B. Desu, J. Mat. Res. 8 (1993): 339), these temperatures are still higher than those which have used with success to deposit PZT by MOCVD techniques of the prior art.
Therefore, a process is desired for the formation of thin films of PZT and related materials, which affords compositional control, provides uniformity of the thin film material over large areas, and achieves a high degree of conformality on the substrate structure, as well as a high deposition rate. The deposited material should also be free of pinholes, since in capacitive and many other devices, pinholes will result in a shorted, useless device.
For thin film PZT and related materials, precise and repeatable compositional control is required in order to produce films of high quality. Physical deposition methods (e.g., sputtering, evaporation) of thin film deposition are deficient in this regard, as are traditional approaches to MOCVD involving the use of bubblers.
Turning to ferroelectric PZT, it is generally recognized that many of the electrical properties can be improved by replacing A or B site species with cations of a higher oxidation state. This is typically referred to as donor doping. In specific cases improvements in leakage resistance, fatigue and imprint have been attributed to donor doping. Improved leakage resistance is observed for donor doping and is believed to be a result of compensation of native and impurity acceptor defects. Improvements in fatigue have been reported for doping with yttrium (Y), (Kim, J. H.//Paik, D. S.//Park, C. Y.//Kim, T. S.//Yoon, S. J.//Kim, H. J.//Jeong, H. J., `Effect of Yttrium Doping on the Ferroelectric Fatigue and Switching Characteristics of Pb(Zr.sub.0.65 Ti.sub.0.35)O.sub.3 Thin-Films Prepared by Sol-Gel Processing`, INTEGRATED FERROELECTRICS, (10), 1995, pp. 181-188), lanthanum (La), (Shimizu, M.//Fujisawa, H.//Shiosaki, T., `Effects of La and Nb Modification on the Electrical-Properties of Pb(Zr,Ti)O.sub.3 Thin-Films by MOCVD`, INTEGRATED FERROELECTRICS, 14, 1997, pp.69-75), niobium (Nb), (Tuttle, B. A.//Alshareef, H. N.//Warren, W. L.//Raymond, M. V.//Headley, T. J./Voigt, J. A.//Evans, J.//Ramesh, R., `La.sub.0.5 Sr.sub.0.5 CoO.sub.3 Electrode Technology for Pb(Zr,Ti)O.sub.3 Thin-Film Nonvolatile Memories`, MICROELECTRONIC ENGINEERING, 29, 1995, pp.223-230.), and tantalum (Ta), (Choi, G. P.//Ahn, J. H.//Lee, W. J.//Sung, T. H.//Kim, H. G., `Phase Formations and Electrical-Properties of Doped-PZT/PbTiO.sub.3 Films Deposited by Reactive Sputtering Using Multi-Targets`, MATERIALS SCIENCE AND ENGINEERING B-SOLID STATE MATERIALS FOR ADVANCED TECHNOLOGY, 41, (1), 1996, pp.16-22.). Significantly enhanced imprint resistance was demonstrated by donor doping with Ta using a sol-gel thin film fabrication process (W. L. Warren, D. Dimos, G. Pike, B. Tuttle, and M. Raymond, "Voltage shifts and imprint in ferroelectric capacitors", Appl. Phys. Lett., 67 (6), (1995), pp. 866-868.).
Doping is straightforwardly achieved using deposition processes such as sputtering and sol-gel. To achieve doped PZT by MOCVD requires identification of suitable precursor chemicals that decompose to the desired product and do not undergo undesirable interactions during delivery and transport to the substrate. Few examples of doping by MOCVD exist in the prior art, and as will be described, many have never been described to date. The most common dopant, La, has been deposited by MOCVD to provide PLZT films as reported by Van Buskirk, P. C.//Roeder, J. F.//Bilodeau, S., `Manufacturing of Perovskite Thin-Films Using Liquid Delivery MOCVD`, INTEGRATED FERROELECTRICS, (10), 1995, pp.9-22.). While Nb doped PZT has been demonstrated by Shimizu et al., they used tetraethyl Pb, Zr(tertiary t-butoxide).sub.4 and Ti(isoproxide).sub.4 for Pb Zr and Ti precursors, respectively, and Nb(ethoxide).sub.5 as a Nb precursor (Shimizu, M.//Fujisawa, H.//Shiosaki, T., `Effects of La and Nb Modification on the Electrical-Properties of Pb(Zr,Ti)O-3 Thin-Films by MOCVD`, INTEGRATED FERROELECTRICS, 14, 1997, pp.69-75). The disadvantage of this approach is that tetraethyl Pb is toxic and relatively hazardous due to its high vapor pressure at room temperature and the lack of a suitable sensor to warn of its presence. Furthermore, the Nb precursor Nb(ethoxide).sub.5 is not compatible with safer Pb precursors, such as a Pb(tetramethylheptandionate).sub.2, used in the present application. It must be emphasized again that the discovery of a compatible set of well-behaved precursors for each doping application is essential.
There is therefore a pressing need in the art for new approaches to the deposition of novel thin film materials of such type, and for next-generation piezoelectric and ferroelectric materials applications.